Monocytes and macrophages can be activated with a variety of cytokines to kill tumor targets in vitro. A variety of animal studies have also suggested that activation of monocytes and macrophages in vivo can bring about tumor responses in selected model systems. A previous study at the Biological Response Modifiers Program suggested that human peripheral blood monocytes can be activated in vitro and adoptively transferred into the peritoneal cavity of patients with peritoneal carcinomatosis and that this approach to the treatment of cancer may have limited efficacy. In this study we will administer recombinant human granulocyte macrophage colony-stimulating factor (rHuGM-CSF) intraperitoneally to patients with disease limited to the peritoneal cavity. Previous studies in mice have suggested that administering GM-CSF in this fashion will result in the recruitment of large numbers of monocytes and macrophages into the peritoneal cavity. If this can be accomplished in humans, the hypothesis that monocytes and macrophages can bring about tumor responses in humans can be tested. In three separate parts of this study patients will receive either GM-CSF alone, GM-CSF with interferon-gamma, or GM-CSF with interleukin-2 (IL-2). All drugs will be administered intraperitoneally. Four patients have been enrolled and treated so far. We have seen substantial increases in the number of monocytes and granulocytes in the peritoneal fluid of three patients. No tumor responses have been observed. Thus, monocytes can be recruited to the peritoneal cavity, but additional patients will be required before activation of monocytes in vivo and antitumor activities in vivo can be assessed.